Re: HTTP router point-of-view concerns

[Willy Tarreau <[e-mail]>]

On Sat, Jul 13, 2013 at 08:28:13PM +0000, Poul-Henning Kamp wrote:
> In message <[e-mail]>, Willy Tarreau writes:
> 
> >Yes. Client picks a random session ID with the highest 16 bits = 0,
> >sends the request to the first server. The load balancer overthere
> >puts the DC ID and the local server ID in these bits and sends this
> >ID back to the client along with the response. When upon a subsequent
> >request the client is directed to a different DC, the information
> >about the location of the client's context is found and the context
> >can be retrieved.
> 
> Thanks!  Now I understand the goal.
> 
> It's unclear to me if it is a good idea to make part of the
> session identifier, or if it should be a separate field, but
> I really like the basic idea.

I think it makes sense to have it part of the session ID because
anyway it's very likely that the server side will use the complete
version to try to improve unicity (which is still not guaranteed).

Willy